永磁同步电机在线参数辨识方法研究

石建飞 戈宝军 吕艳玲 韩继超

摘 要：针对在永磁同步电机参数辨识过程中，由于“数据饱和”和噪声影响，导致传统的递推最小二乘法存在参数估计误差大和收敛慢的问题。利用改进的递推最小二乘法提高参数辨识的精度和收敛速度，以满足伺服系统在不同工况下动态性能。首先，结合永磁同步电机数学模型，设计了一种折息递推最小二乘辨识算法，通过在传统的最小二乘法中引入“折息因子”增强了算法的灵活性。然后，通过对存在白噪声干扰的永磁同步电机模型进行辨识算法的动态仿真。最后，利用搭建的实验测试平台进行算法的实验验证。仿真和实验结果表明本文提出的折息递推最小二乘算法，在參数辨识过程中降低了旧数据对辨识结果的影响，增强了算法对噪声干扰的鲁棒性，提高参数辨识结果的准确性和实时性。

关键词：永磁同步电机；参数辨识；折息递推最小二乘；数据饱和

Abstract： In the process of parameter identification of permanent magnet synchronous motor， due to the influence of data saturation and noise， the traditional recursive least squares has the problems of high error and slow convergence in the parameter estimation. Using the improved recursive least squares algorithm can improve the identification accuracy and rate of convergence， thus meet the dynamic performance of servo system under different working conditions. First of all， combined with the mathematical model of the permanent magnet synchronous motor， a discount recursive least squares identification algorithm is designed， and the flexibility of the algorithm is enhanced by introducing the "discount factor" in the traditional recursive least square. Then， Dynamic simulation of identification algorithm was finished of the motor with white noise model. Finally， experiments were carried out using the experimental test platform. The simulation and experimental results show that the discount recursive least squares algorithm effectively reduce the influence of old data on the identification results and enhances the robustness to noise interference， and improves the accuracy of parameters identification and real time.

Keywords： permanent magnet synchronous motors； parameter identification； discount recursive least square； data saturation